NextGen AI Skills Safety and Social Value – technical mastery aligned with ethical standards

Where do we see a talent gap? What is the requirement in terms of growing this whole ecosystem? Because when we come and we talk about today, it is the era, this is the most exciting time in the industry because AI is transforming everything. AI is transforming the way the businesses are being conducted. AI is transforming the whole workforce also because it’s not about what you are able to do, but it’s about co -exist with the whole AI ecosystem together. So my name is Sibodh and I’m director of SGPA headquarters. I’ve been part of the industry, I’ve been part of the government for almost 27 years. And being in the space of technology, being in the space of working closely within the startup ecosystem, within the academias, there’s always a gap in opportunity which we have witnessed.

And that’s why this particular topic today is very, very close to my heart in terms of how we ensure the industry move forward, how do we ensure that the AI as a technology can bring transformative changes overall. so I am happy that today you know very briefly today’s discussion will align very closely with the national efforts I am sure all of you when you talk about the NDIAI overall theme some of you have witnessed already that there is a lot of activity around the skilling, there is already 10 lakh AI skill in drive which has been initiated there is already a skill India digital program happening this is a new version of skill India altogether within STPI we have focused on you know a program called STPI skill up and I am happy to in fact announce also here that we are going to soon start the multiple regional hubs for training and ensuring that the training across technologies can happen and we have been joined by a lot of our training partners, the current training partner ecosystem is around 18 training partners across India and two of them in fact are here today with us and three of them are here with us and as we move forward, we’ll add more such training partners and collaborators.

We are calling them partners and collaborators because the aim and the objective is all aligned within the ecosystem of skilling up, right? The SIPI skill up becomes that particular program. Let me introduce our speakers. I’m not taking much time. So it’s my privilege to introduce my first speaker, Professor Dr. Alok Pandey, a professor and dean of UP Jindal University, a very senior academic leader with almost three decades of experience, focus across finance, governance, higher education, and I think multiple implementation within the financial technology space. He also comes with a great perspective on the AI. So let me request Professor Dr. Alok Pandey to come on stage and take the space. Please welcome Professor Pandey with a big round of applause.

A limited audience, but ensure that your applause covers the whole hall also. I’ll also like to introduce and welcome Professor Dr. Jawar Singh Professor Dr. Jawar Singh is also a professor from the Indian Institute of Technology Patna, he is also founder of Kuturna Labs and just we were chatting and he has just briefly told about his successful exit so he is not just the professor who is teaching but he is also practicing the same in the form of his own ideas implementation so we are literally and I am sure we are proud to have you Dr. Jawar Singh please welcome you on the dais let me also introduce Dr. Devinder Singh Deputy Director General of TEC this is the Department of Telecommunications in India Dr. Devinder Singh has spent multiple years in the standards formalization standards ecosystem because you understand the telecom space especially is governed by the standards and these standards are very critical because unless and until because the interoperable ecosystem can only work if each and every device each and every node can be standardized and has to be standardized right So, Dr. Devendra Singh, Sri Devendra Singh represents the government from the Postal Telecom.

So, let me welcome with a warm applause from the audience, Dr. Sri Devendra Singh on the dais, please. I’m also honored to join by Dr. Sarabjot. Dr. Sarabjot Singh Anand, he’s a co -founder and chief data scientist of TATRAS, also the co -founder of Sabath Foundation. I have known, you know, Sarabjot Singh from almost, if I’m not wrong, seven, eight years now. And I’ve seen his passion in the space of AI. It’s not about just what he wants to achieve through his, you know, TATRAS data, but also about how, you know, and I think his work in the space of growing AI talent is well recognized in probably in some regions, especially in the region of Punjab, right?

So, Dr. Sarabjot, thank you for being here. I request and welcome you on the dais of pioneer data science. A big round of applause for him. He has also roots in academia at Warwick and Ulcer. He has a very global perspective in this particular space altogether. Let me introduce our next two speakers or two panelists on this agenda today. Vikas Srivastava, he’s a chief growth strategist of Vincis IT Services Private Limited. Vincis is one of our technology training collaborator and partner of STPI Scalar program. Vikas has almost 16 plus years in enterprise consulting, cloud workforce upskilling. And I think Vikas has a great perspective to share in terms of what is really reskilling requirement today within the whole ecosystem of the AI workforce.

So, with a big round of applause, please welcome Vikash Srivastava. last but not the least let us also give a warm welcome to Kunal Gupta managing director of Mount Talent Consulting you know he has been doing talent advisory he runs his own job search portal he understands he works very closely with industry and has a clear perspective what is the industry requirement and where is the gap so with a round of applause Kunal welcome on the dais as well thank you for everyone and let me you know let me probably switch my place as well so it will be easier for us to start the whole discussion hello yes so I think let me quickly start and I will probably start from my immediate left Dr. Sarabjot you know when we talk about next gen AI AI you And when we talk about next -gen AI as a space, next -gen AI as the whole, from the talent perspective, from the opportunity perspective, what is your perspective?

Briefly, we’ll touch upon each one of you on the defining next -gen AI so that the audience understands very clearly what does next -gen AI really means. So over to you.

So to me, there are two camps here, right? One is the people who want to generate the next wave of AI, and then, of course, they’re the ones that have to use AI to be more efficient in their jobs. Now, for both of them, I think what is very, very important is that they have to be critical thinkers more than any technology as such, because there is, you know, a great move towards outsourcing your thinking to AI, and that’s a problem. We need to recognize that AI is not perfect. We need to recognize that there are certain deficiencies in it. and therefore we have to question what we get from that AI. And if we can get people who can critically think about the problem they are trying to solve and then take risks, I think risk -taking is going to be another very, very important aspect and having a foundational understanding of what is possible today with AI and what is not possible today with AI.

Because if we don’t recognize the deficiencies and start to regard AI as an oracle that always tells us the truth, we are going to get into trouble. So these are very, very important aspects apart from of course technology. Thank you.

To Devender Singh, your perspective on the next -gen AI technology in a very brief.

Hello. Next -gen AI, I feel he should have a strong expertise in AI and he should have skills to solve the real -world problems also. And he should adapt to new technologies also. He should be able to work in research. He should be able to work in different sectors also. And above all, I feel he should be aware of the regulations. Thank you. in the sector and in AI also. Thank you.

Thank you. Yes.

Yeah, hello. So to me, actually, the NextGen AI should not only the AI, NextGen should be aware of the AI algorithms, but basically they can make products or solution with customer facing. And they should understand not only the algorithms, but the way those algorithms are mapped onto the hardware. To me, a grounding, solid grounding of hardware, solid grounding of computer science, or even the engineering domain is must, actually. Thank you.

Yes. Professor Alok, sorry for my mistake in pronouncing your name wrong. Yes.

Thank you. I think the next gen AI is largely a T -shaped thing. You need to be domain specialists, deep domain specialists. You need to be fluent in AI skills, whatever software, hardware, etc. you are looking at. And then you should be able to understand red teaming and containment. So, if you have these three, then probably we will be able to solve most of the problems we face in India today.

Please, Kunal.

I think your question is very important. What do I understand or what do we understand with next gen AI? You know, next gen AI is the infrastructure. It’s not for intelligence like you currently have this infrastructure wherein we are able to express our views and they go out to the world. you know next generation of AI is like this infrastructure meant to multiply our intelligence our reasoning our research our values our creativity our judgments and what the future holds for us you know we are going to see a new wave of new materials and for a very long time we haven’t seen any major materials coming apart from the basic alloys that we have been using the process changes which are going to come about in the next generation with the use of the next generation AI the generation of models you know we talk about many things about differentiation in the society from a digital divide to this new edge AI divide but it could also at the same time help us reach out to the inclusive society in general with vernacular languages you know multiplying and extrapolating the reach of what a common normal common man can do earlier they were dependent on languages like English you but with the expansion of the next -gen AI platforms, tools, local vernacular languages wherein you can speak and give instructions to the computer in Hindi, in your local languages, get access to data, knowledge.

Like I said, you know, you could just build anything. We have seen this with a tool called TikTok, you know, a tool about 10 -12 years back which started. And it created a wave of influencers, otherwise a language or a platform meant only for the English and the literate. You know, went on to the masses. So I think next -gen AI, like I said, is just in one word an infrastructure of intelligence, multiplying our ability to think and, you know, make judgments in the future as well. Thank you.

Very well said. It is the infrastructure level intelligence which can be, which has to be created and which defines the next -gen AI. And carrying forward the same thought, I’ll ask Vikas to share his opening remarks on the next -gen AI.

Thank you. So I think most of the important aspect has on you. I think we’ve covered the panel. What I wanted to add is for me, the next -gen talent combines three important things. First is technical mastery. Second is ethical judgment. And third is real -world problem -solving capabilities. So we need people who understand, as I said, you know, we should know, the people should know where AI fits and where AI doesn’t, right? So I think this is the most important thing which I wanted to add. Thank you.

I think for the audience, it is important to understand that when we talk about next -gen AI and we talk about next -gen AI talent and the next -gen AI talent gap, right, there is, we got a clear perspective right from a critical thinking, right, going to the level of not just the, you know, opening up the layers of the AI, but from the perspective that one has to start thinking about the new ways and new layers in which the AI technology is having an impact. But there’s materials which, you know, Whether it is you know the infrastructure intelligence again which we talked about. Whether it is a foundational knowledge and foundational you know algorithm which we talked about.

The next gen AI talent gaps exist everywhere. And accordingly you know I think I will ask Sarabhjotji now to probably talk something on specifically. You know from your perspective of both as TATRAS and Sabud Foundation. You have seen the whole AI evolution. And you have seen the gaps which have been there. And you have tried to fill the gaps already. So my question to you is you know when you talk about the evaluative you know evaluation of the fresh AI talent. How do you what is your approach? Because that approach will lead us you know in terms of ensuring that you know how will this whole space will grow up right. So you are opening Ramon on that here.

Sure thank you. So you know when we look at talent today. What we assess is their problem solving skills. We look at. you know how keen are they on learning themselves so have they taken control of their own agency in learning in the future because what’s happening today and I’ve seen this over the years right a lot of students are very focused because they want to get a job they are focused on learning libraries you know even in 2018 when we started Sabudha Foundation because we found there was a huge gap in AI skilling here in India we found that till we got them to program a neural network they felt they weren’t doing anything right and now of course it’s LLMs everybody wants to learn lang chain and that’s about it but they have to understand the foundations the foundations if they are weak we are going to do interesting things but are not going to do amazing things and so the focus has to be on building a strong foundation increasing their curiosity in terms of what they are doing and getting them to think about how they can be creative in the solutions that they are engineering for their customers.

Now, in Tataras, we work with startups in the US and develop their AI for them. Now, to do that, somebody mentioned domain being very important. And what we are constantly training our folks to say is understand the problem from the customer’s perspective. Right? It’s not just about algorithms. When you create a solution, a successful solution is going to be one that solves the problem. It doesn’t matter what technology you use. And that is a key differentiation between the training that we provide and what is available otherwise in terms of just skilling on libraries. Right.

Thank you. And I think, you know, for all the people outside sitting here, the most important part, I think, as Sarojotji said, and probably any one of you can add also, whenever you feel like, you know, curiosity is one part. Right? Because curiosity to our human mind. adds that element of learning and when the curiosity is there there comes a creativity and once you have this curiosity combined with creativity then only you can understand the customer problem if I am not wrong you can understand the customer ecosystem it’s just about the customer ecosystem from the perspective where you make money but when you talk about social impact because social impact even the people who are getting benefited from the technology they might not be directly paying you but you are creating great amount of social equivalence outside there so it becomes important from their perspective and when we combine these three and we map that with the AI which is such a powerful language right now in terms of technology I think the solutions which you see outside are just a few examples of what really can wonders can be created when you know you bring these three elements together right so I think in similar lines you know I will ask Dr. Devinder Singhji because he comes from his background on the whole telecom space right and today we talk about native AI telecom infrastructure.

When you talk about native AI, they not be just AI ready, but they need to also bring in AI in their own operations. When I say AI readiness, it’s all about the scale, the kind of compute, kind of technology, kind of infrastructure they need to create. But how do they approach from a standards perspective when you see? When you see from the standards perspective, because you see future. You are looking into 6G as a standards. What is the role of an AI in terms of standard creation? And what is the role in terms of technology when standards are getting defined? So from that angle, your thoughts are on the same.

The present telecom engineers, they are very strong in networking. But the future network that the 6G would be coming, it will be more dependent on the AI. The present technology, 5G, is in that case the AI is add on that. But in 6G, each and every component has got AI inbuilt in that only. So at present, the planning is done in a static way. Components are selected and then the effect is seen. But in 6G, it will be self -learning type of thing. So the engineers would be required to know machine learning. And the present cases, whenever there is some fault and alarm is generated, an engineer is supposed to take corrective action. But in 6G, it would be happening, AI will predict what kind of fault can come and it will take corrective action on its own.

And at present, most of the decisions are taken at the central level only. But in 6G, the intelligence will be distributed at the edge also. So the decision will be taken at a distributed level. So the engineer must be able to, plan everything. Thank you. in considering that the distributed decision will be taken. So in that case, as far as the standards are concerned, standards for 6G are being finalized. They are not final, but it is already decided that each and every component will be having AI. In addition to that, you were talking about the standards. Since in TEC, in the telecom engineering center, we have already published some standards on AI. So the telecom engineers should also be aware of the standards which they are supposed to follow for implementing of an AI.

At present, the telecom engineers are using AI, but in future, the telecom engineers will design and operate and use it. Most of the decisions will be taken by AI. The human will only supervise only. Thank you.

Thank you. I think when we talk about telecom space, I think two or three critical things, which probably, you know, is of interest to the audience, as Devendraji talked about, you know, we talk about the AI, agentic AI and the agentic AI in taking care of the operations part. And I think from a skill perspective, it is important that when you look into the agentic AI ecosystem, you need to go deeper into the particular technology or particular sector, because each industry gives its new challenges and problem even for agentic AI ecosystem, right. And when you talk about the infrastructure readiness, right, it is important that today, the whole telecom sector is one of the sector and I think I’ll come back to you on the on the perspective of how is telecom sector creating robustness.

Right. And I think it is important that today, the whole telecom sector is one of the sector and I think it is important that today, the whole telecom sector is one of the sector and I think it is important that today, the whole telecom sector is one of the sector and I think it is important that today, the whole telecom sector is one of the sector and I think it is important that today, the whole telecom sector is one of the sector and I think it is important that today, the whole telecom sector is one of the sector and I think it is important that today, the whole telecom sector is one of the sector and I think it is important that today, the whole telecom sector is one of the sector and I think it is important that today, the whole telecom sector is one of the sector and I think it is important that today, the whole telecom sector is one of the sector and I think it is important that today, the whole telecom sector is one of the sector and I think it is important that today, the whole telecom sector Next, I think I’ll touch upon again, you know, I think to Professor Jawar, you know, when you look into the layer of the hardware and below, right?

Because when we talk about the AI and we talk about six layers as has been spoken about across the spectrum, the most promising and most important layer is also about not just applications, but also about the hardware, which is powering up the whole AI, you know, the need and speed

All right. So, I mean, this is quite interesting because very rarely people talk about how those algorithms runs, how those AI models runs actually. So honestly, if I say these models are very expensive, expensive in terms of not cost, but in terms of power, I will say. And cost is obviously associated with it actually. So, if I say a simple, if I take a simple example, the power requirement of if I take a very basic NVIDIA processor, it consumes around 500 to 700 watt. But if I say the same processor, I mean, not, I mean, our human brain is also having a very beautiful processor. I can say it can compute a lot and it just consumes 20 watt power.

So, you can see there is a huge gap between. The processing capabilities, the most instead of the processors that we have and the. most cognitive processes that we all have is there is a lots of gap actually so the gap need to be bridged so there are lots of research is going on in this domain we usually call the neuromorphic computing or brain inspired computing where these algorithms can be mapped at the hardware in a efficient manner another example i can give you the deep seek when it comes and when it first time pops up or surfaced in the market actually the nvidia’s stocks i mean slumped down quite severely and the reason was that their processor was quite efficient actually that was the only thing so people think okay we may do the same thing in a more efficient way so there is a we need people basically they not only think from the algorithm perspective but they also think from the hardware perspective hardware security also i will add here one more term hardware security because the ai can be weaponized and can also be used for neutralization purpose.

So the hardware play a very crucial role. Algorithms are okay. So we need people basically they understand not from the algorithm but all the way down to the hardware implementation. How your implementation is secure, trusted and the reliable. So I hope No, thank you. Thank you.

You touched upon the element of the security and I think when AI comes into play, the cyber just not the security elements of the algorithms but also the important thing which has popped up is the biasness and the robustness and the fairness of the algorithms. So I am sure some of you will talk about that as a gap from a talent perspective and how do we skill and reskill the people who can be used for this particular filling the layers across the layers they can be probably be more coming into the ecosystem. So with that I think I will ask Dr. Alok. when you look into from a university academia perspective today we talk about population scale AI implementation and I think when we talk about population scale AI implementation it is not about the whole critical thinking and the thinking part also need to be changed and hence academia needs to be geared up to create that kind of curiosity and learning from the perspective of students so what is your take from the industry and academia when you see a gap.

How are you gearing up your students from a perspective of scale of AI perspective and from that particular element?

We have large contracts. Say I have to do an M &A valuation, an M &A due diligence. And, you know, competition commission has asked me whether I should go for this merger acquisition or not. I am a lawyer. How do I do it? I can use a generative AI software for that. I can do money laundering prevention, not just spam prevention like, it’s very effective spam prevention by, you know, Airtel and all. But I can do money laundering calculations and identify which transactions work in which manner through a generative AI. So we need to develop products in these lines. The second thing I would say that safety and security of these products. How are we going to look at, you know, the safe usage?

Now, there’s a term which has come up. You know, this is called a coming wave. Mustafa Suleiman has written a book, The Coming Wave, and everybody uses. This is the coming wave. And the wave is going to drown all of us if safety and security is not there. Every young person who uses AI needs to understand what is red teaming and what is containment. I should kill my technology if it doesn’t work in my favor. Right. And finally, domain integration. So AI healthcare, AI law, AI education, AI finance. So all these levels basically need to be understood by educational institutions. If you ask me another question that how do we scale it up, then I’ll of course later speak on that.

But I’ll tell you that we need to really work out an infrastructure. We need to work out on academic strength. We need to have large number of trained faculty members. We need to have MOUs with Western countries. All companies are based in either China or Europe or America. And the universities are generating a lot of trained resources. And Indian universities need to move forward in that direction. So I basically feel that yes, there is a huge gap today. And we need to really answer these gaps through not just viable funding from government, but also from industry.

I tend to. Partly agree in terms of, you know. the length and the breadth of the AI ecosystem has changed dramatically everywhere but you know when we look into the Indian talent right and I strongly believe because I have been in this industry for very long now and kind of you know energy we are seeing in this particular 10 arenas hall and probably the conferences happening other side there is a huge talent which has popped up now and they are generating very good solutions and today India from a solution producer perspective they are not just doing something at the application layer or the agentic layer but they are also looking into the foundational layer and that’s why we have saw the launch of the recent LLMs also right and when we look into let’s say the launch of the Sarvam LLM or other LLMs it’s very clear people are trying to see there is a lot of data available in our country and this data needs to be understood and as I think as you talked about if you take just one particular sector called law and justice right and there is one start one company here Lex Leges and I was interacting with them yesterday they have understood this problem they have created their LLM on particular you know not the large language model but they have approached this problem with the same mentality of LLM right and hence they have been able to ensure exactly what you are describing as a problem you know how do you create a solutions for that and it works on Indian data it works on the Indian contract laws it works on the Indian you know past judgments so that is the need of our now you as an entrepreneur if you have entrepreneurial mindset as from a perspective of audience sitting here or as somebody who wants to get into this particular as a workforce you need to clearly have the idea about each in every domain whether it is health as you talked about or whether it is law and justice has its own set of challenges and problems right and when there are challenges and problems with right skill and right talent you can actually approach and be very successful and we see this as a you know leapfrog moment for each one of you From the industry perspective also So taking that thought forward I’ll ask Kunal Kunal you have been talking about Skill gaps Especially working with students and working professionals From your platform perspective Whatever you are seeing From your own job portal and job placements perspective What is your take again In terms of most commonly seen The certain abilities Which is required Which probably should be seen by each one of them In a short term What are the skills they need to fill in Whether it is Learning to coexist With the LLMs outside there Whether it is Learning to do the coding Whether in the AGI As Professor Alok talked about Or creating new Machine learning algorithms What do you see as a typical problem In a short term Where talent has to be ready for that

I think I see the problem threefold when it comes to skill gap, specifically in a dynamic country like India, wherein we are living across many generations across the country. You know, a generation which is far ahead in the future, a generation which is far behind in terms of development, capability and education as well. The biggest skill gap that I see right now is the application. And more importantly, how do we define a problem? You know, what we do is we have this mentality out of whatever ecosystem that we have built, we just start copying others in terms of this is the trend, so we need to go for this trend without really understanding how to define the problem first.

Define the problem is about 50 % of the solution achieved in itself. Define the problem in any sphere that the person is in. You know, like Dr. Saab said about different usage in different fields, whether it is healthcare, whether it is… whether it is law, whether it is… is agriculture for that matter, which is catering to such a huge population in our country. Who would have thought of hydroproning, you know, producing such huge results without the application of soil, no dependency on weather, and you can create your own environment for creating absolutely green vegetation, you know, in the best of atmosphere without germs and without the usage of application of pesticides as well. So coming back to your question, skill gap again is going to be defined sector specific.

Different sectors are going to have different specific gaps at different specific application levels. When it comes to industry, again, what is the solution that I am providing or we are providing as a company, you know, our aim is to develop an employability intelligence layer. You know, how do we define skill gap, basis, what kind of jobs are coming from the market, basis, the jobs, what is the current skill set of the candidates, we have a scientific gap analysis in terms of what is missing. it’s not that we have a very nice application tracking system we do a recommendation algorithm using a lot of ai the aim is you know in my view the aim is not to exclude people or reject people using ai when it comes to skill gap analysis the aim is to show them that this is what is missing this is what has to be developed it is not rocket science that can’t be developed you take a course of one month three months six months or you do it while working in another job role while coming to your ambitious job role you know it takes time nothing is built in a day but more importantly i think a more bigger gap that i see right now which is going to come as a huge pressure on the educational setups whether it is at a university level or at a school level you know any which ways we keep talking about the fact that india’s syllabus is very uh you know it is not aligned to the industry it is not aligned to the industry it is not aligned to the industry it is not aligned to the industry it is about 20 year old we don’t update our syllabuses you know it takes six committees to take five years, seven years to come up with new curriculums.

By the time the new curriculum is implemented, it has already gone obsolete from an industry perspective. I think in the last six months, the speed of growth of AI that we have seen is going to put the maximum pressure on the policy makers of the country, specifically those catering to one is the core foundational education, the higher skilling education, and more importantly, the industry skilling which has to be needed to ensure that people understand why productivity is needed, how productivity is done. Students need to understand that most industry needs output. We need production. We need results. We are not, you know, industry cannot always bridge the gap. And in India, you know, I’ll have to say it, whether it is an MSME or whether it is large industries, everybody has done their bit in terms of scaling those whom they select.

And, you know, today the success that we see in conferences like these is a lot of people. People who have grown through the industry and how industry has scaled people up. Correct. Colleges will need to ensure that AI education is for all. You know, application of AI for all. Output will increase. Output will lead to, you know, more analysis in terms of how to improve the output production. Production leads to more research. Research is going to lead to more efficiency in production. And it’s a loop. You know, currently the application is going to lead to higher output in my view. Higher output in terms of what an engineer can do in eight hours of work.

What a company can do in terms of per year revenues. Whatever models can do. Whatever processes can do. And based on that, it’s a regular running cycle. We can’t sit in a relaxed manner right now. Specifically in this changing world right now.

I’ll just add to what Kunal has said. We need to de -bureaucratize education today to a great extent. In fact, we brought in this concept of institution of eminence. And I’m happy that I’m part of an institution of eminence. where we can create our own curriculum. You know, curriculum velocity is so high that you can’t give a command to faculty members to teach a particular course, especially in technology. And especially when you’re talking about integrating with a particular domain when the faculty has to work with other specialists, identify something and the needs change frequently. You know what has happened? It’s not just that AI technologies have changed. The consumer and the user has started demanding change.

For example, if you look at the crop insurance, the crop insurance idea basically means that I should have satellite pictures and I should have an understanding as to whether a crop failed or not. And this is done best using AI. And if I need to train my agriculture college students, you know, who study in large agronomy institutions, I need to have a quick delivery of the curriculum. Sadly, we don’t have that. We don’t have expertise on those. Thank you. So if you de -bureaucratize curriculum, allow more autonomy to institutions who are into technology at least, or technology applications, we’ll have a much bigger national good at hand.

And I think the start has already happened with the NEP, if I’m not wrong, right? The whole focus on the national education policy and the initiatives around that has been giving more autonomy and speed towards defining the curriculum. So I tend to see this as a problem which was there. And there’s already a lot of work has happened now, right now. I think if you talk about, I think, when you were at probably a globally level and you were probably, you know, from a Warwick experience, right? You would have seen these changes there. And do you see this coming back to India right now in that similar speed?

I don’t, unfortunately, right? So at Warwick, we actually have the Jaguar Land Rover research labs on campus, right? And we were interacting with them. Even 14 years ago, we were looking at tracking, the cognitive load on a driver. as they drove a vehicle to understand whether we need to take some preventative action before he causes an accident, right? Now, of course, we are saying we don’t need a driver. So times are changing very quickly. I think the one thing when we started Sabud, we realized that curriculum is falling short. Academics are not equipped to deal with the change that’s happening. Even HR folks, when we look at it from an industry perspective, our HR folks are not evolving quick enough to evaluate candidates the right way, right?

So what we did in Sabud was, we said the centerpiece of our training is what we call a passion project, where we get students, we are training them in AI machine learning and technology, but we are getting them to think about how do I solve a problem of social impact? Right? And then we are giving them mentors. who are from Tatras and from other organizations that are actually creating AI solutions for the global north, as they say, right? And so now the students are getting mentorship. And I think the key thing we are missing today, which is shocking for a country our size, we have companies with lots of technologists that have no choice but to keep up with technology innovation.

At the same time, these people have to be trained to give back more. If we can get every person to be evaluated or valued based on how much they give back to others, then we can pair students with mentors in industry and get them to get the skills that no curriculum can give, right? Because you really need problem -solving skills that are existing outside of academia. Of course, academia. Academia has great depth, and therefore they have to be part of that. And so as Subodhji was saying, we’ve got to bring academia and industry closer together and solve this problem. It’s not going to happen from one side alone.

No, I think, thank you for sharing your thought on the I’ll just take you know, professor, sorry, Vikash first and then you on, please go ahead.

Specifically for this, so in that way basically from the curriculum point of view at least, I just want to make this a small caveat actually related to this curriculum updates at least the centrally funded technical institutions are not a problem at all. Even they are free of all those things actually. If I have to start a new course from the next semester, I’m free to run. So such kind of restrictions for curriculum updates, at least as far as I know, CFTIs are not bound to that actually and they are quite okay.

It was not only for CFTIs because India is 1 .4 billion people, right, and majority of it are in tier 2. My basic problem is state technical institutions. The talent which comes from state technological universities, which is the best talent. And these people need scholarships. These people need multilingual support. And these teachers also need training. You know, and, you know, there’s a very large layer for the state institutions because education being both a center as well as state funded thing. And we are in a quagmire where, you know, new regulators are coming in, old regulators are falling and we need to identify how to do it. But my basic thing was not CFTIs. The central funded institutions are much better off.

But still, you know, the amount of manpower you need for developing AGI kind of systems. And it is yet to see just a matter of five years. We’ll see how this hypothesis works, whether we are able to generate something in artificial general intelligence. I think all of us will have to contribute towards this transformational change right from academia to the industry, to the policymakers like us. It becomes important. We understand the speed is not required. But. to develop the solution, speed is required in terms of how the solutions get developed by virtue of doing right things, right?

And I think to Vikash, I think Vikash, you have seen, because you have been coming from the AI learning space, right? So my only question to you is you would have seen the conventional way of doing AI education in past and how that has changed today, right? Are we still looking into conventional classroom mechanism of making the AI learning or as you know, probably what Sarab said, it’s not about learning but practicing it while learning, right? So what’s your input around the same?

So in my view, conventional or traditional trainings, they focus heavily on theory, mathematics, model architecture, you know, those foundations are important but from an industry readiness, we require additional three layer. First is, you know, problem, I know, applied problem solving. So, learner must work on real data set. Or, you know, they should focus on the domain -specific knowledge. Or they should work with the deployment scenarios. Second is, you know, the production exposure. So, knowing how your model move from your notebook environment to real scalable or, you know, secure, you know, systems. And, you know, how the production happens. And last is.

So, I think when we talk about the classroom learning and we talk about the learning about the mathematics. How do you see the, you know, the new tools and technologies being used for training? For example. you know are there any examples probably you can quote some examples we have seen that students are now able to not just see the typical you know learning of the classroom but what other tools and technologies they’re being exposed so the learning gets you know increased this the learning speed of learning becomes faster?

So basically in in our sector you know we are utilizing ai to assess the skill gaps so now there are tools who based on the you know participant profile is able to assess the learning gaps and recommend adaptive learnings so which eventually help you know increasing the employability outcome that’s that’s so so this is how the ai is helping today.

Great i think while we you know we are probably somewhere almost towards the end but we have one more set of questions but just to keep the audience anybody wants to have one quick question please can somebody bring the mic to them? Can somebody please bring the mic to them ? Right so I think we’ll I wanted to go one more round of questions but just to keep the interactiveness because the audience is also limited I don’t want you guys to get bored about what we are speaking so anybody can probably ask one or two questions?

Thank you hello everyone namaskar just quickly you speak in Hindi you tell your name my name is Vikram Tripathi I am from a village in Prayagraj and the upcoming elections are the panchayat elections I am going to participate in them there is a district panchayat election there is a district panchayat member of 25 villages so if I win the election then in the first year the AI tools or softwares which are available which are the three sectors where I should use them and secondly is it possible that private companies, CSR funds Thank you.

One bias index is produced depending upon matrices. For one bias, I can use a number of matrices also. Result of all the matrices are clubbed to find the bias index for one particular parameter. Then, a system can have bias due to many things and different bias indexes are clubbed to find one fairness index. The fairness index ranges from 0 to 1. If it is 1, then it is considered fair. If it is 0, it cannot be used. But in practice, the fairness index will be from 0 to 1. Then it will depend upon the user also. If he wants to have how much fairness in the system. If the system is used to suggest what song would you like to hear, then some bias may be accepted.

If system is supposed to identify whether the person is the soldier is enemy or our own. then no bias can be accepted so that can be used for the by the deployer and those matrices or the framework we have suggested it can be used by deployer the developers also the engineers who are involved in developing those systems those people can also test their models if it is fair or not and it can be used by the regulators also the regulator may say the government may say for such sector the system should be tested and it should have at least this much fairness level similar to fairness we have got one standard for robustness also which can be used to check if the system gives consistent results in different situations

Great and I am sure these standards are available in the public domain they are not draft stages.